The utility of multiple synthesised views in the recognition of unfamiliar faces

The current data (SJDDMB_Experiment1_2. xlsx) from two experiments are reported in two tabs. In both tabs data are averaged across trials and participants are organized by rows. For Experiment 1, the first two columns (A & B) represent participant demographics (Age & Gender). The following four columns represent a mean accuracy (i.e., proportion correct) score for each of the four experimental conditions reported in the paper. Scoring a correct answer as 1 and an incorrect answer as 0 and then averaging scores across trials gave a mean accuracy. The experimental conditions comprised computer-generated views created from a single original image (Synthesised Views), the target image alone (Original image), a single test image that displayed a face at the test angle (Test image) and multiple photograph views (Photographic Views). The second measure reported is a confidence-accuracy (CA) score. The CA score was calculated by multiplying accuracy (negatively scored for incorrect answers so 1 = correct and −1 = incorrect) by the confidence score (1: “Not at all confident”, 7:“Extremely confident”, minus 0.5) giving a score between −6.5 and +6.5 in 13 equal steps. Data are again averaged across trials and reported for each experimental condition. The data are organized in a similar fashion for Experiment 2 with the addition of a Distractor column. This column indicates if a participant was shown a distractor face before test.

Results derived from these data are published in the Quraterly Journal fo Experimental Psychology at http://dx.doi.org/10.1080/17470218.2016.1158302 . These experiments explored how introducing multiple views of an individual enhanced later identification. Some of these multiple views were taken as photographs and some were artificially generated.